ML20210H258
| ML20210H258 | |
| Person / Time | |
|---|---|
| Issue date: | 05/28/1997 |
| From: | Apostolakis G Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-3020, NUDOCS 9708130321 | |
| Download: ML20210H258 (9) | |
Text
--,~~nnnnnevnsn
~
Geor,so Aposto1ckis - 5/28/97
~&O d it/ h n g ADVISORY COMMITTEE G
JOINT MEETING OF THE SUBCOMMITTEES ON INDIVIDUAL PLANT EXAMINATIONS AND ON PROBABILISTIC RISK ASSESSMENT MEETING MINUTES - JUNE 11-12, 1996 ROCKVILLE, MARYLAND INTRQDUCTION The ACRS Subcommittees on Individual Plant Examinations (IPE) and Probabilistic Risk Assessment (PRA) met on June 11-12, 1996, at 11545 Rockville Pike, Rockville, MD, in Room T-2B3.
of this meeting was to discuss the IPE Insights ReportThe purpose (June 11) and the NRC staff research programs related to PRA (June 12).
The entire meeting was open to public attendance.
Mr. Michael T.
Markley was the cognizant ACRS staff engineer for this meeting.
l The meeting was convened at 8:30 a.m. each day and recessed at 3:27 p.m. on June 11 and adjourned at 12:05 p.m. on June 12.
ATTENDEES ELCRS G. Apostolakis, Chairman D. Miller, Member (6/11 only)
I. Catton, Member W. Shack, Member M. Fontana, Member M. Markley, ACRS Staff T. Kress, Member R. Sherry, ACRS Fellow W. Lindblad, Member Princloal NRC Sceakers M. Cunningham, RES J. LaChance, SAIC*
A. Camp, SNL*
J. Lehner, BNL*
S. Dingman, SNL*
G. Parry, NUS-Haliburton*
M. Drouin, RES T.
Pratt, BNL*
J.
Forester, SNL*
J. Taylor, BNL*
W. Hodges, RES J. Wreathall, WreatWood Group
- Brookhaven National Laboratory (BNL), Sandia National Laboratories (SNL), Science Applications International Corporation (SAIC), NUS-Haliburton, and The WreatWood Group
{}
l are NRC contractors to the Office of Nuclear Regulatory Research (RES).
b 9 pt - / llU C
~
o p4, _ L. - t I!IH EIll!!H.il:11!!M 3,m n,_a
_,s.s 9708130321 970811 DESIGt;rrsp op;an,g PDR ACRS 3020 PDR
,,,I M b _
f C ol)
Joint IPE/PRA Subcommittee 6/11-12/96 Minutes Industry /Public Presenter B. Christie, Performance Technology i
i A complete list of meeting attendees is in the ACRS Office File, and will be made available upon request.
The presentation slides and handouts used during the meeting are attached to the office copy of these minutes.
JUNE 11, 1996 Inc udaction Dr. George Apostolakis, Chairman of the joint Subcommittees, convened the meeting and stated that this was a two day Subcommittee meeting and the purpose of the meeting was to discuss the IPE Incights Report (draft NUREG-1560) on June 11 and to discuss the NRC research program related to PRA on June 12.
Dr. Apostolakis stated that the Subcommittees had received a request from Mr. Bob Christie of Performance Technology to make a presentation entitled, " Performance-Based Testing of Standby Equipment."
He noted that this document and another entitled,
" Turbine-Driven Pump Surveillances: 'What Do We Learn,'" were previously provided to the Members.
Presentation by Mr. Christie Mr. Christie reviewed the topics for his discussion including the elements of a system reliability program, parameters for performance-based testing, guidelines for testing, and practical considerations of testing and data collection.
Significant points made by Mr. Christie include The three elements of conditional probability include e
standby availability, probability of start, and probability of run, The intent of standby testing is to duplicate actual demands o
at certain calendar-based intervals.
Performance-based parameters are measurable, objective, quantitative criteria that provide for flexibility in the monitoring process to establish cost-effective corrective
- action, Parameters used in the testing guidelines are based on how o
often equipment is actually demanded, the probability of
Joint IPE/ PPA Subcommittee 6/11-12/96 Minutes success during actual demands, the importance of equipment, and the money available, Practical considerations include effective system o
engineering and " smart" testing through a "living" process to monitor and adjust testing.
l Dr. Apostolakis questioned the definition of availability versus i
probability to start.- Mr. Christie stated availability meant that the equipment was not tagged out of service.
He stated that his numbers indicating probability to start were based on actual demands.
l-Dr. Catton expressed the view that this approach may not represent good engineering practices and may not represent good statistics.
Mr. Christie stated that the data represent relative values and not absolute values.
He stated that the plant system engineers update the information every month, t
Drs. Apostolakis and Kress asked whether the IPE was useful and whether plants should have PRAs.
Mr. Christie stated that the main value of the IPE was educational and that plants should have Level 3'PRAs.
NRC Staff Presentation Mr. Mark Cunningham, Chief, Probabilistic Risk Analysis Branch (PRAB), Division of Systems Technology (DST), RES, provided an overview of the IPE Insights Program, its scope and limitations, and described how the program elements fit together.
Ms. Mary Drouin, RES, led the discussion for the staff.
NRC contractors from BNL and SNL also participated and provided supporting discussion.
Mr. Cunningham stated that the IPE Insights Report was developed
-to gain perspectives on plant designs and vulnerabilities.
The perspectives were derived from the IPE results from 75 licensee IPE submittals covering 108 operating reactor units.
The staff defined the scope of its--review which included consideration of Level 1 and Level-2 PRA results for full power operation and internal events.
The staff did not evaluate the quality of IPEs nor did it consider external events (including fire) or other plant modes such as low-power and shutdown operations.
-Ms. Drouin and cognizant NRC contractor personnel reviewed the following program areas:
Core damage frequency (CDF) perspectives Containment performance-perspectives e
e-Impact on reactor safety: Vulnerabilities and improvements
_IPEs with-respect to risk-informed regulation e
Joint IPE/PRA Subcommittee
-4 6/11-12/96 Minutes e
Additional IPE perspectives The staff summarized the objectives of the program related to reactor safety, reactor and containment design, the importance of
{
)
operator actions, current IPEs as compared with state-of-the-art
)
PRAs, and the Commission's safety goals.
The staff also considered improvements identified as a result of the Station Blackout Rule, and compared with insights gained from NUREG-1250,
" Severe Accident Risks:
An Assessment for Five Nuclear Power Plants."
The staff summarized its observations and conclusions for each of the objectives.
Significant points made during the discussion include:
only a few licensees identified vulnerabilities, but all e
identified safety enhancements or plant improvements.
Most improvements were directed at reducing risk for station blackout.
For BWRs: CDF does not decrease for newer BWR designs; CDF variability was primarily due to design / operation differences, modeling assumptions and data; and failure to depressurize is dominant contributor to CDF.
l For PWRs: CDP variability driven primarily from plant-specific support system design and operational differences; support system failures are dominant contributors to CDP for most plants; and plants with better feed and bleed have lower CDFs.
For BWRs: conditional containment failure probability (CCFP) for Mark-I containments have higher early structural failure versus later containments which are largely driven by dry-well shell melt-through; bypass and containment isolation failure is generally not important for BWRs; and the large volumes of Mark-III containments are partially responsible for lower CCFPs.
For PWRs:
CCFP low early structural failure for ice condenser containments appears to be driven by analysis assumptions rather than design features; bypass, especially steam generator tube rupture, is important to early structural failure for both ice condensers and large dry and subatmospheric containments; and isolation failures are significant in a number of large dry containments.
For BWRs Important human actions are required for manual depressurization, containment venting, initiation of standby liquid control (SLC), and alignment of containment to suppression pool cooling.
s
i Joint IPE/PRA Subcommittee
-B-6/11-12/96 l
Minutes For PWRs: Important human actions are required for switchover to recirculation mode of emergency core cooling, feed and bleed, and depressurization and system cooldown.
Concerns were identified in some IPEs relative to human reliability analysis, common-cause failures, data analysis, and success criteria.
However, Level 1 PRA methods were generally sound.
Perspectives in IPEs were consistent with NUREG-1150 except G
that the likelihood of early containment failure is higher for PWRs with ice condenser containmenth than those with large dry containments.
Subca--ittee Questient and Co-rnts Drs. Apostolakis.and Kress raised questions regarding the consideration.of uncertainty in IPEs.
Dr. Apostolakis questioned the relative contribution of aleatory (random) uncertainty versus
-epistemic (state-of-knowledge) uncertainty.
The staff stated that some plants did uncertainty analysis but most did not.
It is very difficult to differentiate aleatory uncertainty from epistemic uncertainty in the IPEs.
Dr. Apostolakis asked about the impact of modeling and assumptions on the results of analysis reported in the IPE Insights Report.
The staff stated that it did not fully examine the acceptability of modeling and assumptions provided by licensees but noted that there was a lot of variability between some IPEs.
The staff stated that modeling and assumptions were a factor and that it had asked for additional information from some licensees.
However, major differences were noted in site-specific design features.
Dr.-Shack questioned the probability assumed for operators taking required actions such as those needed to depressurize the vessel.
The staff stated that there were some differences in the HRA modeling but reiterated that most differences were attributed to plant differences.
Dr. Apostolakis asked whether the staff attempted to validate the assumptions and models or just accepted them.
The staff stated that most were accepted as reported.
Mr. Lindblad asked whether most licensees used state-of-the-art practices in their'IPEs or generally accepted industry practices.
Dr. Apostolakis~ asked whether there were any models that were considered unacceptable.
The staff stated that it did document the approach used by licensees with regard to current stata-of-the-art PRAs but indicated that it looked very closely at the consistency of application for each model selected.
Joint IPE/PRA Subec9mittee
-6 6/11-12/96 Minutes JUNE 12. 1996 IntroductioD Dr. Apostolakis, Chairman of the joint subcommittees, convened the meeting and stated that he had a conflict with portions of the mater!al to be presented and announced that Dr. Kress would serve as Acting Chairman for the meeting.
Dr. Kress noted that this was the second day of the two-day Subcommittee meeting and the purpose of the meeting was to discuss the NRC research program related to PRA.
He also noted that the Subcommittees had requested that the staff focus its presentation on those rescatch activitiea related to human reliability analysis.
Dr. Kress noted that no written comments or requests for time to make oral statements were received from members of the public for this portion of the meeting.
NRC Staff Presentation Mr. Mark Cunningham, Chief, RES, provided an overview of the NRC l
research program for PRA.
Ms. Mary Drouin, RES, led the discussion for the staff.
NRC contractors from BNL, NUS-Haliburton, and The WreathWood Group also participated and pr*vided supporting discussion.
Mr. Cunningham summarized the approach and rationale for the research activities related to PRA.
He noted that plant evcnts and precursor analysis of those events have indicated that failures in human performance are dominant sources of risk and that " errors of commission" are a key ingredient in all major accidents.
He indicated that humans occasionally operate plants outside the rules and good practices, that physical processes are not understood by operators, and that operators fail to recognize or refuse to believe the evidence.
He added that existing methods do not always appropriately represent / treat human performance.
Mr, Cunningham stated that the objectives of the NRC research program is to improve HRA methods to estimate the contribution of human performance to plant risk.
Ms. Drouin and cognizant NRC contractor personnel reviewed the following program areas:
ATHEANA - A Technique for Human Error Analysis e
Review of HRA methods used in IpE submittals Significant points made during the discussion include
,=
Joint IPE/PRA Subcommittee 6/11-12/96 Minutes ATHEANA:
The purpose of the ATHEANA project is to develop a method that allows more realistic assessment and representation of human contribution to plant risk.
It is based on analysis of risk mignificant operational experience and understanding of behavioral sciences.
e Ai'HEANA provid%9 a discipline to better understand the untato Antion the error mechanism, and the error forcing
(
cota$xt.
The behavioral scles.ces perspective provides insights for e
underctanding the logic of human errors, combinations of systems behaviors and human factors.
These insights can be used to derive principles to anticipate when errors are more likely - although successes and failures are never certain.
ine ATHEANA model has been developed using 'he following
+
task descriptions:-
Task 1:
Familiarization with the PRA model including event sequences and modeling assumptions Task 2:
Identification of potential human _ failure events and associated unsafe actions Task 3:
Identification of most probable /significant causes of the unsafe actions Task 4:
Refinement of HFE definitions Task 5 Estimation of event probabilities including the likelihood of plant conditions and conditional human error probabilities HRA Methods in IPEst Methods used in IPE HRAs include the following:
-e Methods used by the Electric Power Research Institute (EFRI) include the Human Cognitive Reliability Method (HCR), Operator Reliability Experiment), and Cause-Based Decision Tree Method Technique for Human Error Rate Prediction (THERP)/
Accident Sequence Evaluation Program Method Success Likelihood-Index Method Modified THERP Method Operator Reliability Characterization Method The staff evaluated each of these methods and had the e
following generic concerns:
I For time-based methods: there was limited systematic I
analysis of " times required to perform actions;" the basis
Joint IPE/PRA Subcommittee 6/11-12/96 Minutes for selecting " times available for action" was unspecified and varied by cutset; and there was no consideration of uncertainties in " times available for action."
General conclusions were that all methods have limitations, e
all depend on skill of the user, and applications of the modified THERP Method were found not to meet the requirements of Generic Letter 88-20, " Individual Plant Examination for Severe Accident Vulnerabilities" Subca-=4 ttes Questions and C=--anta Drs. Apostolakis and Kress asked about the extent to which the IPEs considered errors in human performance.
The staff stated that it did not address them in a detailed way.
The ATHEANA project is intended to make improvements in this area, particularly with regard to evaluating errors of commission (EOCs).
Dr. Apostolakis asked about the separation of " safety culture" from " management and organization."
The staff stated that it was
-focusing on the impacts of safety culture.
Dr. Apostolakis pointed out that taere might not be a definition of safety culture that can be agreed on.
He expressed the view that safety culture could be two cultures, an informal and a formal one.
With regard to the likelihood of plant ccaditions that are not modeled in the event trees, Dr. Apostolakis stated that efforts should be given to areas which are not modeled and that the assumptions of those that are modeled should be examined more closely.
Followun Actions The Subcommittee requested the staff to provide an overview of the IPE Insights Report at the full ACRS meeting, June 12-14, 1996.
The Subcommittee requested this presentation also cover
-the. major points covered during the June 11 Subcommittee meeting regarding this matter.
The Subcommittee informed the staff that the briefing to the full committee does not-need to-cover the NRC research program on PRA and noted that a followup Subcommittee meeting to discuss the ATHEANA Project would be appropriate in about a year.
~
RDr. Apostolakis_ requested copies of various NUREG documents (NUREG/CR-6093, NUREG/CR-6265, and NUREG/CR-6350) associated with the ATHEANA project.
4
Joint IPE/PRA Su'scommittee 9-6/11-12/96 Minutes Backaround Material Provided to subca==4ttee for this Meetina 1.
SECY-96-051, " Status of IPE and IPEEE Programs," dated March 8,
1996 2.
Letter dated April 1, 1996, from B. Christie, Performance Technology, to G. Apostolakis, ACRS, Subj: Request to present papers on performance-based testing tc the IPE/PRA Subcommittee 3.
Draft position _ paper dated hay 8, 1996, Subj
" Performance-Based Testing of Standby Equipment," B. Christie 4.
Draft paper for International Conference on Probabilistic Safety Assessment and Management, June 24-28, 1996, " Turbine l
Driven Pump Surveillances,'What Do We Learn?'" B. Christie S.
Report dated March 8, 1996, from T. Kress, Chairman, ACRS, to S. Jackson, Chairman, NRC, Subj "Use of Individual Plant Examinations in the Regulatory Process" 6.
Letter dated April'10, 1996, from J. Taylor, EDO, to T.
Kress, Chairman, ACRS, Subj: "Use of Individual Plant Examinations in the Regulatory Process" 7.
-Report-dated: April 23, 1996, from T. Kress, Chairman, ACRS, to S. Jackson, Chairman, NRC, Subj "Probabilistic Risk Assessment Framework, Pilot Applications, and the Next Steps to Expand the Use of PRA in the Regulatory Decision-Making Process" 8.
-Draft NUREG 1560, " Individual Plant Examination Program Perspectives on Reactor Safety and Plant Performance" NOTE:
Additional details of this meeting can be obtained from a transcript of this meeting available in the NRC Public Document Room, 2120 L Street, N.W. Washington, D.C. 20006, (202) 634-3274, or can be purchased from Neal R. Gross & Co.,
Inc. Court Reporters and Transcribers, 1323 Rhode Island Avenue, N.W.
Washington, D.C.
20005, (202) 234-4433.
I